Use and Perceived Value of Health Information Exchange: One Public Healthcare System's Experience

Health information exchange was adopted and accepted rapidly in a large public healthcare system. Its use varied by patient characteristics, clinical settings, and type of insurance.
Published Online: November 22, 2013
David C. Kaelber, MD, PhD, MPH; Rehan Waheed, MD; Doug Einstadter, MD, MPH; Thomas E. Love, PhD; and Randall D. Cebul, MD
Objectives: To describe health information exchange (HIE) use and providers’ perceptions of value in a public healthcare system using a commercial electronic health record (EHR).

Study Design: Observational study of HIE implementation and cross-sectional provider survey.

Methods: We identified characteristics (age, gender, race/ethnicity, insurance type, comorbid conditions) and the care setting (primary care; emergency department [ED] or inpatient care; or specialty care) for patients with and without HIE. Associations between patient characteristics and HIE were examined using a multivariate logistic regression. Provider perceptions were assessed via confidential survey.

Results: During its first 14 months, 11,960 HIEs occurred among 9399 patients. Rates of HIE use were 13/1000 visits overall (20/1000 in primary care, 36/1000 in the ED/inpatient setting, and 5/1000 in specialty settings [P <.001]). Patients with HIE were older, more often female, African American, had more chronic conditions, and more often had Medicaid or Medicare insurance (P <.001). HIE was used least among commercially  insured (odds ratio, 0.78, 95% confidence interval,0.73-0.83, compared with uninsured). Among the 18% (74/412) of survey respondents, 93% “disagreed/strongly disagreed” that obtaining consent was difficult and 97% reported no patient refusals. Respondents “agreed/strongly agreed” that HIE fostered more efficient care (93%), saved time (85%), decreased laboratory (84%) and imaging (74%) use, and 15% stated that HIE prevented an unnecessary admission.

Conclusion: Early HIE use varied by care setting, patient characteristics, and insurance. Providers perceived HIE acceptable to patients, and helpful in avoiding redundant testing and unnecessary hospitalizations. Lower HIE use among commercially insured patients reinforces concerns that financial incentives may inhibit adoption.

Am J Manag Care. 2013;19(11 Spec No. 10):SP337-SP343
Health information exchange (HIE) adoption has been lower than expected, but few studies document actual use and provider perceptions. In this study of HIE use in a public healthcare system:
  • Adoption occurred over several months—highest in the emergency department and inpatient settings and lowest among sub-specialty providers.

  • Providers reported little difficulty obtaining patient consent for HIE.

  • Providers reported the benefits of HIE-averted testing, reduced hospitalizations, and improved efficiency.

  • HIE occurred more often among patients who had more medical conditions and had Medicare or Medicaid insurance.

  • HIE occurred least often among commercially insured patients, indicating that financial incentives and market competition may inhibit widespread use.
As electronic health record (EHR) use becomes widespread, the associated increase in the use of electronic health information exchange (HIE) could add significant value to thehealthcare environment and reduce costs. Conceptual studies estimate that the United States could save billions of dollars through HIE.1-5 However, the adoption of HIE has been slow, with current estimates suggesting only 10% of US hospitals use HIE.6 The Health Information Technology for Economic and Clinical Health (HITECH) Act7 provides billions of dollars to encourage “meaningful use” of EHRs in the hope that the resulting savings will blunt increases in healthcare expenses.8 Stage 2 meaningful use criteria mandate some HIE use,9 with increased HIE use requirements expected in stage 3.10

Extracting significant value from HIE use requires that EHRs become ubiquitous so that digital data are available to all entities involved in a patient’s clinical care. Once these digital data exist, robust HIE systems must allow Health Information Privacy and Accountability Act (HIPAA)- and HITECH-compliant, easy, meaningful exchange of data. Also, HIE must be readily adopted and used. With significant HIE use, immediate value should be derived from decreased duplicative testing, identification and reduction of avoidable redundancies in care, and higher overall levels of safety, quality, and value. Some studies suggest the value of HIE in specific settings—including emergency departments (EDs) and human immunodeficiency virus care.11-15 While demonstrating reductions in tests and admissions and higher-quality care, these studies do not examine HIE use across a healthcare system with a broad spectrum of patient characteristics, including multiple comorbid conditions, different care settings, and different insurance types. We described HIE adoption and its perceived value to providers in a public healthcare system using a vendor-based HIE integrated in a commercial EHR.


Setting and Subjects

We implemented a real-time, pull, EHR vendor-based (Care Everywhere, Epic Systems Corporation, Verona, Wisconsin) HIE in our academic, public provider, integrated tertiary care system in northeast Ohio. The healthcare system includes a 731-bed hospital with 17 outpatient sites, over 400 staff and 350 resident physicians, and a level 1 trauma center. Annually, the system has approximately 900,000 outpatient visits, 100,000 ED visits, and 30,000 inpatient admissions. Payer mix includes 21% uninsured, 24% Medicaid, 18% Medicare, and 37% commercial insurance. Northeast Ohio now has 5 other health systems using the same vendor’s EHR and the vendorbased HIE includes over 1500 sites nationwide, conducting over 1.25 million exchanges per month. Exchanges reported here were with a single large nonprofit system of 10 hospitals and owned health centers, which was the only other system on the vendor-based HIE during the study period.

HIE functionality is fully integrated with the EHR. The HIE uses demographic information (name, gender, address, and telephone) to identify out-of-system information for the patient in real time. If a unique match is found, after attestation of signed patient opt-in consent, the other system’s EHR information is pulled into our EHR. The requestor is informed if no unique match is found.

HIE implementation started in the ED in November 2010 and deployed systemwide in January, 2011. Providers in all care settings (ED, inpatient, primary care, and specialty care) had HIE access after systemwide deployment and were taught using e-mails, EHR tips, and optional in-person training. Demographic and context data on HIE patients during the first 14 months of implementation were abstracted from the EHR and compared with all other patients in the period. Five to 7 months after HIE implementation, we administered a survey (eAppendix) to all providers who had used the HIE at least once.


We identified all HIE requests made for adult patients (aged >18 years) seen in ED/inpatient, primary care, or specialty care. We defined an HIE request as the first attempt on the encounter date to obtain new or additional information through the HIE. We tracked monthly HIE requests. For all patients with and without an HIE request we abstracted the care setting (ED/inpatient, primary, or specialty care, defined using EHR visit encounter department) in which the HIE requests or visit occurred as well as patient characteristics (age, gender, race/ethnicity [White, African American, Hispanic, other/unknown]), insurance (commercial, Medicare, Medicaid, uninsured), and comorbid conditions. Insurance was taken as the insurance at the time of HIE use or the most recent visit for non-HIE patients. The 11 chronic medical conditions investigated on the past medical history or problem list (with associated International Classification of Diseases, Ninth Revision, Clinical Modification codes) were: arthritis (711, 712, 714-716, 720, 726), coronary artery disease (414), cancer (140-239), congestive heart failure (428), chronic obstructive pulmonary disease (490-496), hypertension (401-405), liver disease (751), stroke (430-438), psychiatric disease (295-301, 308, 309, 311), diabetes (250), and renal disease (585). We excluded patients less than 18 years old and HIE use in a nonclinical department. Patients withmultiple HIE uses were assigned to the care setting of first use. Characteristics of HIE and non-HIE patients were compared using χ2 for categorical and t-tests for continuous outcomes. The rate of HIE use for each department setting was calculated as the number of encounters where HIE was used divided by total departmental encounters. We performed a multivariate logistic regression to examine the independent relationship between patient characteristics and the use of HIE across all care settings. SAS version 9.3 (SAS, Cary, North Carolina) was used for all analyses.

Cross Sectional Survey

After our HIE had been available for 5 months, we surveyed all providers who had used the HIE at least once. The survey was constructed and distributed using RedCap.16 Nonresponders were sent 2 reminder e-mails. The 14-item survey asked about HIE usability, patient consenting, and perceptions of cost-relevant issues including efficiency, time spent, impact on potential hospital admissions and use of laboratory and imaging tests, other potential benefits, and barriers to use. The MetroHealth System Institutional Review Board approved the observational study and survey.


Longitudinal Observational Study

In the first 14 months of use, 11,960 information exchanges occurred for 9399 unique patients. We excluded 929 patients under age 18 years and 35 patients seen in nonclinical care settings; the remaining 8435 comprise our HIE sample. Most patients (8202) had only 1 HIE request during the 14-month period, 205 had 2, 24 had 3, and 4 had 4 or more. During the same period, 143,567 adult patients had no queries through the HIE. Among HIE requests, 82.8% successfully identified the patient’s information at the queried outside institution. Figure 1 shows the monthly rate (per 1000 encounters) of HIE requests by care setting (ED/inpatient, primary care, and specialty care). HIE requests increased quickly and then plateaued across all care settings—rising faster and indicating faster adoption in primary care than in ED/inpatient or specialty care. After 14 months, an average of 750 HIE requests were made per month.

Table 1 compares HIE and non-HIE patients. HIE use was associated with age (50 vs 45 years), gender (female: 61.1%), African American patients (42% vs 34%) and those with more chronic conditions (2 vs 1.2). Patients with HIE were less often commercially insured (20% vs 32%) and more likely to have Medicare (24% vs 15%) or Medicaid (24% vs 22%) (all P <.001). HIE patients in the ED/inpatient group were also less frequently uninsured (39% vs 53%). The results are also shown in a multivariate logistic regression analysis (Table 2).

Cross-Sectional Survey

Of 428 unique providers who used the HIE at least once during the first 5 months (51% of the 839 providers in the MetroHealth System), 412 had a healthcare system email address and were e-mailed a survey. Eighteen percent (74/412) of surveys were returned across all clinical settings. Among survey respondents, 93% disagreed or strongly disagreed that they had difficulty obtaining HIE consent and 97% indicated that no patient had refused consent. Eightysix percent reported successfully completing HIE requests most of the time. From HIE requests, providers reported viewing laboratory, imaging, and other tests more than summary information or clinical notes (93% of the time vs 84% of the time vs 85% of the time).

PDF is available on the last page.
COPD Compendium
Dermatology Compendium
Diabetes Compendium
GI Compendium
Immuno-oncology Compendium
Lipids Compendium
MACRA Compendium
Oncology Compendium
Rare Disease Compendium
Reimbursement Compendium
Rheumatoid Arthritis Compendium
Know Your News
HF Compendium
Managed Care PODCAST